The present invention relates to a moving picture coding apparatus which is utilized for a visual telephone and a television conference.
In compressive coding of a moving picture, a moving picture coding technique in which orthogonal transformation and quantization are combined with each other is widely used. FIG. 4 is a block diagram showing a coding technique such as ITU-T H.263 and ISO/IEC MPEG-4. An input picture 10 is processed at a pixel block (16×16 or 8×8, for example) unit. In a subtracter 12, a difference between the pixel block of the input picture 10 and a pixel block of a predicted picture 20 is calculated. Discrete cosine transform is applied to the said difference by a DCT (discrete cosine transformer) 13. In a quantization circuit 14, an output from the DCT circuit 13 is quantized by a quantization step size 24 determined in a coding control section 23. An output from the quantization circuit 14 is converted into a code by a variable length coding circuit 21, and is output via a buffer 22. On the other hand, an output from the quantization circuit 14 is inverse-quantized by an inverse-quantization circuit 15, and inverse-discrete cosine transform is applied thereto by an inverse-DCT circuit 16, and the predicted picture 20 is added thereto for every block in an adder 17, and it is stored in a frame memory 18. A local decoded picture block stored in the frame memory 18 and the input picture 10 are supplied to a motion vector detecting circuit 11, and a motion vector is detected. Based on the local decoded picture from the frame memory 18 and the motion vector detected by the motion vector detecting circuit 11, motion compensation is conducted by a motion compensation circuit 19, and the predicted picture 20 is output. The coding control section 23 determines the quantization step size 24 from the remainder (generated code content) of the buffer 22, and outputs it to the quantization circuit 14 and outputs frame skip control information 25, and cuts off an input of the input picture 10.
Next, a state in which the coding control section 23 determines the quantization step size 24 to be used in the quantization circuit 14 based on the remainder information of the buffer 22 will be explained by referring to FIG. 5. At a step 41, the buffer remainder is compared with a threshold value T1, and when the buffer remainder is greater than the threshold value T1, at a step 42 a constant number C1 is added to the quantization step size 24, and by making the quantization rough, the generated coding content is made less. When the buffer remainder is less than the threshold value T1, at a step 43 the buffer remainder is compared with a threshold value T2, and if the buffer remainder is greater than the threshold value 2, at a step 44 a constant number C2 is subtracted from the quantization step size 24, and by making the quantization fine, the generated coding content is made greater. Next, clip processing (steps 45-48) is conducted so that the quantization step size 24 calculated in this manner is put between preset minimum quantization step size and maximum quantization step size.
As a method of setting a minimum value and a maximum value for the quantization step size, a method shown in JP-P111768/1988A can be raised. Although the minimum value and maximum value to be set are defined as a standard of a coding method such as H.263 and MPEG4, there is a case where a value to be taken is further restricted for picture quality control. For example, by setting the maximum value of the quantization step size to be smaller, it becomes possible to suppress picture quality deterioration of each frame in a picture or the like, which moves hard. However, in this case, to suppress generated information content, it is necessary to use a frame skip or the like together for means other than a change of the quantization step size. In this case, the coding control section 23 generates frame skip control information 25 based on the remainder of the buffer 22.
However, in the above-described conventional moving picture coding apparatus, since a frame skip occurs due to the temporary increase of information content in case of compression to a low bit rate, there is a task that a frame rate becomes unstable.